Abstract Segmentation of tumors in medical images is not only of high interest in serial treatment monitoring of “disease burden” in oncologic imaging, but is also gaining popularity with the advance of image guided surgical approaches. Magnetic resonance images
are widely used in the diagnosis of brain tumors. In this article, an automatic tumor detection and classification system is presented, which focuses on the structural study on both tumorous and normal tissue. The proposed system consists of the following steps: (i) pre-processing, (ii) feature
extraction using an enhanced texton co-occurrence matrix and (iii) classification. In classification, a fuzzy logic based support vector machine is used to classify the experimental images into normal and abnormal. The obtained experimental results show that the proposed brain tumor detection
approach is more robust than other neural network based classifiers, feed forward neural network and radial basis function, in terms of sensitivity, specificity and accuracy.